Arctic sea ice at 1.5 and 2 °C

In the Paris Agreement, nations committed to a more ambitious climate policy target, aiming to limit global warming to 1.5 °C rather than 2 °C above pre-industrial levels. Climate models now show that achieving the 1.5 °C goal would make a big difference for Arctic sea ice.

How warm is too warm for the survival of summer Arctic sea ice? The long-term goal of the United Nations Framework Convention on Climate Change has been to limit the average global temperature rise to 2 °C above pre-industrial levels. The 2016 Paris Agreement saw a more ambitious commitment made, under which nations agreed to pursue efforts to limit warming to 1.5 °C. Scientists are now working to discover what this change in policy means in terms of real-world impacts. In Nature Climate Change, two studies explore what the Paris Agreement means for Arctic sea ice. Using different climate models, Michael Sigmond and co-authors1 and Alexandra Jahn2 both demonstrate that the probability of an ice-free Arctic Ocean can be minimized by preventing this extra 0.5 °C of warming, but that occasional ice-free summers are still possible1,2.

Arctic sea ice has declined rapidly over recent decades, in large part due to anthropogenic GHG emissions3. The Arctic Ocean is projected to become ice-free in summer by the middle of this century4, unless emissions are substantially reduced. There is therefore much interest in estimating the likelihood of ice-free conditions for warming of 1.5 °C compared with 2 °C of warming.

September marks the low point in the annual cycle of Arctic sea ice, and so ice-free conditions are expected to first appear in this month, if at all. Sigmond et al. and Jahn both report a low probability (2.4% and 2.5%, respectively) of ice-free conditions in September if the global average temperature stabilizes at 1.5 °C above pre-industrial levels1,2. This implies that ice-free conditions would be expected once in every 40 years, on average. By contrast, 2 °C of warming increases the frequency of occurrence considerably, to once in every three to five years (19% and 34%, respectively, in Sigmond et al. and Jahn)1,2.

Broadly similar probabilities, of 2% or less at 1.5 °C of warming and 34–43% at 2 °C of warming, have been found by other studies5,6,7,8. There is therefore agreement across climate models that the probability of an ice-free Arctic in September would be substantially reduced by achieving the 1.5 °C target of the Paris Agreement. Note, however, that although the reduced odds of ice-free conditions at warming of 1.5 °C relative to 2 °C are clear, the exact probabilities should be interpreted with caution. This is because the sensitivity of sea ice to global warming in the real world is highly uncertain9, which makes it difficult to assess whether sea ice is lost at the correct rate in climate models.

The studies by Sigmond et al. and Jahn also help answer questions of avoided impacts. In the context of Arctic sea ice, would limiting warming to 1.5 °C be sufficient to prevent an ice-free Arctic ever occurring? Previous work suggested a very low probability of ever witnessing an ice-free Arctic in the years leading up to reaching 1.5 °C of warming8. The two new studies corroborate this finding1,2. Yet, they also demonstrate the importance of the cumulative risk in years after reaching 1.5 °C of warming. Their works shows, for the first time, that the probability of an ice-free Arctic occurring at least once continues to increase over time after reaching 1.5 °C above pre-industrial levels, even if warming ceases thereafter1,2.

There is a simple explanation for this increasing probability over time: natural variability. On top of the long-term change due to global warming, there is year-to-year natural variability in Arctic sea ice. As the years go by, there is increased chance of an episode of natural climate variability that leads to extremely low sea ice; and so therefore, the probability of at least one occurrence of an ice-free Arctic also increases.

Consider the analogy of the probability of rolling a six with a standard six-sided dice. For each roll of the dice, the chance of rolling a six is 1-in-6 (17%) and the chance of not rolling a six is 5-in-6 (83%). The more rolls of the dice, the greater the chance of rolling a six at least once. Mathematically, the probability of rolling a six at least once is 1 – pn, where p is the chance of not rolling a six and n is the number of rolls. There is a 31% chance of rolling a six at least once in two throws and a 60% chance in five throws. By twenty-five throws, the odds of rolling a six at least once are 99%.

For the probability of an ice-free Arctic, each subsequent year can be thought of as another roll of the dice. The same equation can be used to mathematically predict how the probability of at least one occurrence of an ice-free Arctic changes over time (Fig. 1), by setting p to be the chance of not seeing an ice-free Arctic and n to be the number of years. The probability of at least one occurrence of an ice-free Arctic increases from 2% in a single year, to roughly 20% over a decade, to 50% within 30 years for warming of 1.5 °C. Although limiting warming to 1.5 °C may be insufficient to prevent an ice-free Arctic altogether1,2, it does seem to be sufficient to avoid ice-free conditions in months other than September. This is not the case for warming of 2 °C, for which ice-free conditions in August become possible2.

Fig. 1: Probability of at least one occurrence of an ice-free Arctic.

Instantaneous probabilities (year 1) are taken from the climate model simulations reported in Sigmond and colleagues1 (solid lines) and Jahn2 (dashed lines), for stabilized global warming of 1.5 °C (blue) and 2 °C (red). Accumulated probabilities (years 2 to 100) are derived mathematically (see text). The climate model experiments performed by Sigmond et al. mimic this mathematical relationship1. Differences in probabilities between 1.5 °C and 2 °C of warming (yellow/orange shading) and the IPCC likelihood ranges (dotted lines) are indicated. An ice-free Arctic is likely to occur at least once if global warming stayed at 1.5 °C for 45 years, or within 10 years for 2 °C of warming.

Given that the feasibility of staying below 1.5 °C of warming, or even 2 °C, has been questioned10,11,12, what would the implications of missing these targets be for Arctic sea ice? Current emissions reduction pledges put us on a path to around 3 °C of warming by 210010, which Sigmond et al. calculate would probably lead to ice-free conditions every summer1. If we overshoot 1.5 °C or 2 °C of warming, could the loss of sea ice be reversed at a later date? In this regard, Jahn’s study offers some good and some bad news. The good news is that sea ice recovers if the globe cools, at the same rate per °C as it was lost2. The bad news is that this linearity does not hold for the relationship between global temperature and CO2 concentration2. To cool the climate, say from 2 °C back to 1.5 °C above pre-industrial levels, the CO2 concentration would have to be reduced by more than it rose in going from 1.5 °C to 2 °C of warming.

Sigmond et al. and Jahn both conclude that the probability of an ice-free Arctic is much lower at 1.5 °C than 2 °C of warming1,2. Further, they make clear an important point that has been missing in past work: that the probability of an ice-free Arctic ever occurring continues to increase after global warming ceases. Even if warming can be held at 1.5 °C, there remains a high chance of an ice-free Arctic occurring at least once; unless warming is limited to 1.5 °C and reduced quickly thereafter.


  1. 1.

    Sigmond, M., Fyfe, J. C. & Swart, N. C. Nat. Clim. Change (2018).

  2. 2.

    Jahn, A. Nat. Clim. Change (2018).

  3. 3.

    Notz, D. & Stroeve, J. Science 354, 747–750 (2016).

    CAS  Article  Google Scholar 

  4. 4.

    IPCC Climate Change 2013: The Physical Science Basis (eds Stocker, T. F. et al.) (Cambridge Univ. Press, 2013).

  5. 5.

    Sanderson, B. M. et al. Earth Syst. Dynam 8, 827–847 (2017).

    Article  Google Scholar 

  6. 6.

    Iversen, T. et al. Earth Syst. Dynam. Discuss. (2017).

  7. 7.

    Ridley, J.K & Blockley, E. W. Cryosph. Discuss. (2017).

  8. 8.

    Screen, J. A. & Williamson, D. Nat. Clim. Change 7, 230–231 (2017).

    Article  Google Scholar 

  9. 9.

    Niederdrenk, A.L. & Notz, D. Geophys. Res. Lett. 45, 1963–1971 (2018).

    Article  Google Scholar 

  10. 10.

    Rogelj, J. et al. Nature 534, 631–639 (2016).

    CAS  Article  Google Scholar 

  11. 11.

    Meinshausen, M. et al. Nature 458, 1158–1162 (2009).

    CAS  Article  Google Scholar 

  12. 12.

    Millar, R. J. et al. Nat. Geosci. 10, 741–747 (2017).

    CAS  Article  Google Scholar 

Download references

Author information



Corresponding author

Correspondence to James A. Screen.

Ethics declarations

Competing interests

The author declares no competing interests.

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Screen, J.A. Arctic sea ice at 1.5 and 2 °C. Nature Clim Change 8, 362–363 (2018).

Download citation

Further reading


Quick links

Nature Briefing

Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.

Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing